Herpes simplex virus (HSV) has often been suggested as a suitable vector for the nervous system due to its neurotrophic lifestyle and its ability to remain in neurons for the lifetime of the cell. However wild type HSV is highly pathogenic and must like most viral vectors be disabled in some way. The pathogenic effects of HSV result from lytic infection with the virus and therefore the use of HSV as a vector requires the development of strains carrying mutations that disrupt the lytic cycle whilst allowing the establishment of asymptomatic latent infections. HSV vectors have previously been produced and tested in vivo by the deletion of the essential immediate early (IE) gene ICP4 (Dobson et al., 1990 and Chiocca et al., 1990), which must be complemented for growth in culture. ICP4 is required for transcriptional activation of the viral early and late genes in lytic infection. Thus, a virus lacking this gene can readily establish latent infection of cells but cannot grow lytically.
Mutations have also been made in non-essential genes such as the IE gene ICP0, the IE gene ICP6, tyrosine kinase (TK), US5 or VMW65, all of which are required for full pathogenicity in vivo but are dispensable for growth in culture (reviewed by Coffin and Latchman, 1996). These types of mutation provide the added advantage that the deletion need not be complemented for growth in culture which has been shown previously to occasionally result in reversion of the non-pathogenic phenotype to a wild-type phenotype by homologous recombination between the virus and the complementing sequences in the cell-line during growth. However in each of these cases, mutation of the non-essential gene does not completely prevent virus replication since high titre inoculation will overcome the block to replication in vivo.
We have tested, as vectors, HSV mutant strains deleted for the ICP34.5 gene--the so-called neurovirulence factor--which is absolutely required for neurovirulence in vivo, but is again unnecessary for growth in culture (Chou et al., 1990). Mutations in ICP34.5 provide a subtle mechanism by which HSV can be disabled. ICP34.5 is thought to prevent the usual host response to a productive infection in neurons, which results (in the absence of ICP34.5) in cell death and thus the limitation of the infection to initially infected cells. ICP34.5 is thought to over-ride this response and allow full lytic replication to occur. Thus in the absence of ICP34.5, if a disabled virus were ever to re-establish a productive infection for whatever reason, the ICP34.5 mutation would ensure that the protective host response limited virus replication to a small number of cells.
To test the possibility that ICP34.5 deleted herpes viruses might be developed as vectors for the nervous system, we inserted a lacZ construct into a non-essential gene (UL43) of the HSV1 ICP34.5 deletion mutant strain 1716, and inoculated mice via the footpad route (for delivery to dorsal root ganglia (DRGs)) and intracranially. LacZ activity (as assayed by X-gal staining), could be seen in a limited number of cells in the DRGs and brain respectively (unpublished observations). These results indicate that HSV strains carrying inactivating mutations in ICP34.5 are suitable for use as gene-delivery vectors. ICP34.5 deletion mutants could thus provide the basis for further development as novel and safe gene-delivery vectors for the nervous system.
However, it is unlikely that viruses carrying a single defect will be considered safe enough for eventual human use. Added safety and the possibility of higher titre inoculation might be achieved by the deletion of an essential IE gene providing an absolute block to replication (and which must thus be complemented in culture), together with ICP34.5. Previously HSV vectors disabled by removal of essential IE genes and used in vivo have been deleted for ICP4 as this single deletion absolutely prevents replication and provides the greatest reduction in the levels of other HSV gene products. However the other IE genes (ICP0, 27, 22 and 47) are still produced and of these the essential ICP27 is highly cytotoxic probably due to its secondary role of preventing the splicing of pre-mRNAs in favour of translation from the mainly unspliced herpes RNAs. We therefore speculated that removal of ICP27 (to be complemented in culture) might produce a safer and less cytotoxic vector system when combined with deletions in ICP34.5. While a number of ICP27 deletion mutant viruses have been produced and used, for example, to study herpes gene regulation in vitro and the effects of ICP27 on the host cell (Reef Hardy and Sandri-Goldin, 1994 and Rice and Knipe, 1990), none has reportedly been tested as a vector for gene delivery to the nervous system in vivo. Furthermore, none of the ICP27 deletion mutant viruses carry a mutation in ICP34.5.